Research and development costs

Transcripción

Research and development costs
L
A
R
E
V
U
Translated from Rev Prescrire November 2003; 23 (244): 782-787
E
Drug costs
Research and development costs: the great illusion
In the first three articles on prescription drug pricing published in our French edition, we examined how
the French authorities have attempted, with little success, to control the prices of new drugs (1-3). French
inadequacies in this area are partly due to the growing power of pharmaceutical companies, against a
background of globalisation, economic liberalisation, and an industry-oriented regulatory framework in
Europe. Now we examine the cost of research and development, a major argument used by drug companies to justify high prices.
● The commonly stated average cost of researching and developing a new drug is 802 million
dollars.
● This figure comes from an institute largely funded by the pharmaceutical industry, and is based
on confidential information also
provided by industry sources.
● It includes the cost of failures
and financial charges, and is the
pre-tax figure. Out-of-pocket
spending actually represents
only half this sum, and the
true costs are halved again
when tax is deducted. A
detailed analysis of this estimate reveals many other
methodological flaws.
● This estimate only
concerns new chemical entities entirely
developed by the
company in question.
● It is based on clinical development costs
which are far higher
than those quoted by
other sources.
● Drug companies
claim this estimate is
the “official” figure.
● Data on the many
drugs that are not fully
developed by the company in question are even less
consistent.
32
• P R E S C R I R E I N T E R N A T I O N A L F E B R U A R Y 2004/ V O L U M E 1 3 N ° 6 9
● The true costs of research and
development must be determined transparently if national
authorities are to make rational
decisions on patents, data protection, and drug pricing.
● Research and development
costs must be viewed in the light
of drug company profits, which
remain the highest of any economic sector.
ll industrialised countries
are now confronted by
escalating health expenditure, largely fuelled by rising
drug-related costs. The latter are
growing with population ageing,
increasing drug consumption
across all age groups (partly due
to advertising pressure), and the
very high prices of new drugs
(1,2).
Drug companies are virtually
free to set their own prices in
industrialised countries (2,3).
Why do the authorities accept
these very high prices when they
threaten the very existence of
national welfare systems? One
frequently cited reason is the
“astronomical” cost of research
and development (a).
This article asks: is the widely
quoted estimate of research and
development costs reliable? Are
there any rival estimates? and
How and why have stated
research and development costs
escalated over the years?
A
Only one information
source since 1979
The alleged cost of researching and developing a new drug,
universally quoted by drug companies and accepted at face value
by governments, journalists and
other experts, derives from a single source, which estimated in
2002 that, on average, a new drug
cost 802 million dollars to develop (4). Before examining the
validity of this estimate, here is
a brief summary of how it
emerged.
An estimate updated every
10 years or so, occasionally
adjusted for inflation.The first
estimation of research and development costs dates back to 1979
(4), when the Tufts Center for
the Study of Drug Development
arrived at a figure of 54 million
dollars per new drug (5). The first
widely quoted estimate of
research and development costs
was published in 1991 (6), by a
team working for the Tufts Center. The method of calculation,
derived from that used in 1979,
was based on a sample of drugs
and on parameters derived from
the Tufts Center database, which
is fed with information from drug
companies(4). In 1991, research
and development were estimated to cost 231 million dollars per
new drug (in 1987 dollars) (6).
This figure was arrived at after
various calculations (see below),
that can be repeated with new
parameters. This was done in
1993 by the Office of Technology Assessment (a former US
public body): by increasing the
capital opportunity cost (see
below) from 9% to a rate varying from 10% to 14%, and by
adjusting for the US inflation
rate during the period in question, the “official” cost of research
and development (universally
quoted, especially by drug companies) became 359 million dollars (in 1990 dollars) (7).
Simple adjustment for the
inflation rate increased this figure to 473 million dollars in 2000,
which drug companies conveniently rounded up to 500 million (8).
Between 1991 and 2002, the
“official” costs of research and
development were simply inflation-corrected updates of the
1991 Tufts Center estimate, itself
based on a sample of drugs whose
clinical development started
between 1970 and 1982 (4).
The Tufts team made a new
estimate in 2001, with new data
collected from a sample of drug
companies. They arrived at an
average figure of 802 million
dollars (in 2000 dollars) for
researching and developing a
new drug (4).
An industry-funded institute. Before examining this latest estimate in detail, the reader
should know that the Tufts Center for the Study of Drug Development is an institute specialising in the pharmaceutical industry. It is affiliated to Tufts University
(Boston), but is financially independent from it. It has a unique
database, furnished by data from
drug companies. The Tufts Center itself is funded by donations
(especially from drug companies),
to the tune of 65%, and also by
sales of products and services
(studies, seminars) (b)(5).
A simple estimate
For all its apparent precision
(802 million dollars) this estimate
of research and development
costs is based, on the one hand,
on data provided by the drugs
industry, and, on the other hand,
on questionable or opaque calculations.
Methodological problems.
This figure of “802 million dollars” is not a precisely itemised,
average real cost, but a complex
estimation. Indeed, it is extremely difficult to attribute specific
research and development costs
to a particular drug: early research
and development activities are
not always linked to a specific
drug, but to several more or less
closely related substances, only
some of which will eventually
be marketed.
Development failures must be
taken into account, as pharmaceutical firms must cover all their
research and development costs
with sales of the few drugs that
are effectively marketed. Estimates of research and development costs therefore usually
include drugs that are abandoned
in the development stage. This
is the case for the “802 million
dollars” estimate.
Another important source of
imprecision is the fact that
research and development lasts
several years, making it even
more difficult to itemise spending. There is a dual risk of overlooking certain costs and of
including spending on other candidate drugs.
To make up for missing data,
the Tufts Center team made
“mean” estimates from its dataset
(see below).
A secret sample of 68 drugs
from ten firms. The figure of
“802 million dollars” comes from
a study of data supplied by 10
drug companies (four of the 10
world leaders, four situated
between the tenth and twentieth place, and two lower down
on the scale) (4). The study
focused on 68 drugs randomly
selected from the portfolios of
the 10 firms. All were developed
by the companies concerned
(none were purchased or sold
under licence). Their clinical
development started between
1983 and 1994, and costs were
counted up to the end of 2001
(27 of the 68 drugs had been
authorised by the end of
2001)(4). The names of the firms
and drugs, and all the other information used by the Tufts Center team, have never been published, and were only provided
by the companies on the understanding that they would remain
confidential.
Weightings derived from a
larger database. Data on the
68 drugs only concerned the
costs of clinical development, i.e.
human and animal studies (4).
The costs of basic and preclinical research were calculated as
a fraction of the total development costs, using data from the
Tufts Center database (containing information on several hundred drugs) (4).
Research and development
costs for the drug sample panel
were then calculated from the
costs of clinical development,
weighted by a mean failure rate
and a mean duration of development calculated from the
database (4).
The figure of “802 million dollars” was arrived at with the following parameters: 21.5% of
drugs that entered phase I trials
eventually received marketing
authorisation; preclinical development (including research
costs) accounted for 30% of all
development costs; the interval
from the beginning of clinical trials to marketing authorisation
was 90.3 months (7 years
6 months); preclinical development lasted 5 years; and the
annual capital opportunity cost
was 11% (4).
Financial calculations. The
notion of capital opportunity costs
was used by the Tufts Center team
to take into account the fact that
research and development
requires several years of investment before receipts start to
accrue. In principle, it corresponds
to what the companies concerned
could have earned by investing
elsewhere rather than in research
and development, on the stock
market for example.
The “802 million dollars” break
down as follows: 335 million
dollars for preclinical development (including 214 million or
63.9% in capital costs); and 467
million dollars for clinical development (including 185 million
dollars or 39.6% in capital costs).
Thus, capital costs represent nearly half (399 million, 49.8%) of
the “802 million dollars” (4).
An estimate with
many question marks
The “802 million dollar”
research and development estimate is highly questionable, for
many reasons.
A single, unverifiable source. The single-source origin of
the estimate undermines its
validity. The study itself is unverifiable, as drug companies themselves provided the data, on condition of secrecy, and solely to
the Tufts Center. In other words,
the Tufts estimate cannot be
independently reproduced. It all
depends on whether or not one
trusts the Tufts Center and the
companies that provided the
data.
Blatant conflicts of interest.
It is obviously in drug companies’
best interests to overstate their
research and development costs.
The Tufts Center’s main clients
are drug companies, and conflicts
of interest are therefore unavoidable. The Office of Technology
Assessment pointed out in a-Pharmaceutical companies prefer to place
the accent on the nobler word “research”.
Yet, in practice, pharmaceutical research and
development is usually based on discoveries (receptors, mechanisms of action, etc.)
made in public laboratories.
Pharmaceutical research and development
include synthesis or extraction of a compound
with a potential pharmacological action;
studies of pharmacological activity (on cells,
tissues, animals, or computer models); formulation of a product administrable to
humans; toxicity studies (on tissues and animals); clinical trials (phases I-III before marketing authorisation; and phase IV after
marketing authorisation); and development
of a large-scale manufacturing process. The
first two stages correspond to research, and
the remainder to development (ref 7).
b- The Tufts Center website cites several
enthusiastic endorsements by industry representatives. For example: “If someone were
to ask me whether they should join the Tufts
Center, I would absolutely recommend it.
We have been a sponsor for a number of
years and I think the kind of information
that the Tufts center provides is just not available elsewhere” (ref 5).
P R E S C R I R E I N T E R N A T I O N A L F E B R U A R Y 2004/ V O L U M E 1 3 N ° 6 9
• 33
1993 that the potential political importance of the Tufts Center estimate might tempt companies to overstate their costs,
with not the slightest risk of being
found out (7). This commonsense remark is even more valid
in 2003, now that “research and
development costs” have become
a major war horse for the pharmaceutical industry.
Unrepresentative.The Tufts
Center estimate is not representative of all new drugs, for
several reasons.
First, it concerns only new
chemical entities (4), which represent a relatively small proportion of new candidate drugs.
According to a drug company
employee, new chemical entities represented only 332
(24.1%) of the 1375 drugs marketed worldwide in the period
1975-2000, while the remainder were new indications, new
forms, new dose strengths, or
new combinations (9). According to the US Food and Drug
Administration, new chemical
entities represented 35%
(361/1035) of drugs marketed
in the United States in the period 1989-2000 (10).
New drugs that are not new
chemical entities are far cheaper to develop. For example, new
indications are generally granted on the basis of clinical trials
alone; range extensions and new
combinations are generally
approved without further clinical trials (c).
Public-sector contribution
underestimated. The sample
used by the Tufts Center team
comprises drugs that were exclusively developed by the 10 companies concerned (4). Yet few
new drugs are now entirely
researched and developed “in
house”, a significant part of the
work being done (or funded) by
the public sector, most notably
in the United States where about
half of all new chemical entities
are discovered by the public sector.
A study published by the US
National Bureau of Economic
Affairs showed, for example, that
14 of the 21 major drugs mar-
34
keted in 1965-1992 had benefited from publicly funded
research and development (8).
A study done by the US National Institutes of Health (NIH) at
the request of the consumers’
association Public Citizen also
showed that public research
played a predominant role in the
research and development of the
five drugs with the biggest worldwide sales in 1995 (aciclovir,
captopril, enalapril, fluoxetine
and ranitidine) (d)(8,11).
This involvement of the public sector is not simply limited to
basic research, but also includes
clinical trial sponsorship and
funding (e).
A very small sample. Readers should note that the cost of
phase III trials, which represents
two-thirds of the costs of clinical
development in the Tufts Center
estimate, was stated for only 33
of the 68 drugs in the sample
(especially considering the withdrawal of some drugs after phase
I or II studies) (4). These 33 drugs
represent a very small proportion
of all drugs marketed during the
12-year period studied by the
Tufts Center: between 1990 and
1999, for example, 284 new
chemical entities were marketed
in the United States (4).
Questionable values.All the
parameters used by the Tufts
Center team are unverifiable and
therefore questionable, including the failure rate, the real duration of research and development, and the percentage of costs
represented by preclinical development.
Research and development
costs are highly sensitive to
changes in these parameters. For
example, “802 million dollars”
falls to 734 million if the global
success rate is 23.5% rather than
the 21.5% used in the Tufts
study (f)(4).
Capital opportunity costs are
also a matter for debate. They
were calculated by the Tufts Center from financial achievements
by drug companies in 1985-2000
(4). This method appears somewhat paradoxical: the more that
drug companies’ share prices
appreciate, the higher the capi-
• P R E S C R I R E I N T E R N A T I O N A L F E B R U A R Y 2004/ V O L U M E 1 3 N ° 6 9
tal costs, and therefore the higher the cost of research and development. In other words, claimed
research and development costs
would be lower if drug companies were less profitable!
Yet capital opportunity costs
are a key element in the Tufts
Center estimate, accounting for
63.9% of preclinical development
costs. Each half-point increase or
decrease relative to the 11% used
by Tutfs would correspond to
25 million dollars (4).
Overstated research period. In its 1991 study, the Tufts
Center calculated capital opportunity costs by assuming an interval of 98.9 months (8 years
3 months) between the beginning of clinical trials and the
granting of marketing authorisation (4). This interval was
reduced to 90.3 months (7 years
and 6 months) in the 2001 study,
because of a marked reduction
in the time required to obtain
marketing authorisation in the
United States (4).
Another Tufts Center team estimated this interval at 87.4 months
for drugs marketed between 1996
and 1998, and 80.6 months for
drugs considered “important” by
the US Food and Drug Administration (35% of drugs during the
period in question) (12). According to this Tufts Center team, relative to 1993-1995, the mean
intervals observed in 1996-1998
corresponded to a 19% reduction in the duration of clinical trials and a 31% reduction in the
time required to obtain marketing authorisation (12).
This latter study also shows
that the interval between the
beginning of clinical trials and the
granting of marketing authorisation is highly variable from one
therapeutic category to another,
undermining the use of the average value. This interval was only
44.7 months for the nine antiretroviral drugs marketed
between 1996 and 1998, and
these were among the most highly priced pharmaceuticals (12).
Only half the stated cost of
R&D is actually spent. It is
important to note that capital
opportunity costs represented
practically half the total figure of
802 million dollars (399 million
dollars). In other words, companies actually spent (out-ofpocket) only half the 802 million dollars.
Tax advantages omitted.
If the Tufts team had consistently used the same accounting logic, they should also have
taken into account specific tax
advantages for research and
development. Indeed, research
and development costs are taxdeductible, contrary to financial
investments (7).
US drug companies can also
deduct 20% of certain research
and development costs from their
tax bill (25% in the United Kingdom) (13,14).
Likewise, 50% of clinical
development costs for orphan
drugs are tax-deductible in the
US (13). Overall, drug companies have a far lower taxation
rate than other industrial sectors
in the United States (26%, compared to 33% on average) (13).
The Office of Technology
Assessment assumed that the cost
of research and development
should be expressed in post-tax
figures (7). On this basis, Public
Citizen estimated that the real
cost of research and development
corresponded to total post-tax
c- These drugs, which are not new chemical entities, and which cost less to research
and develop, should not to be overlooked
(see Prescrire’s annual awards list, for example). Therapeutic advance is not tied to a
novel manufacturing process, or to a new
mechanism of action or compound, but is
simply measured by the risk-benefit ratio
in clinical practice.
d- The websites of Public Citizen
(http://www.citizen.org) and another US
consumers’ association, Consumer Project
for Technology (http://www.cptech.org),
offer many useful documents on the pharmaceutical industry, including research and
development costs.
e- Purely in-house research and development is no longer the rule, as the product
portfolios of the largest drug companies
include an increasingly large proportion of
drugs sold under licence, and which were
largely researched and developed by another company. The world’s largest pharmaceutical company, Pfizer, derived 53% of
its 2002 turnover from drugs sold under
licence (ref 28).
f- Interestingly the proportion of candidate
drugs actually brought to market does not
only depend on the medical or technical failure rate: it also includes commercially motivated decisions (inadequate market niche).
PRESCRIRE
VIEWPOINT
Why the rising costs of pharmaceutical research ?
The costs of pharmaceutical research and
development increased markedly during the
last two decades, without reaching the figure
of “one billion dollars per new drug” predicted by Eli Lilly in 1991 (1,2). This inflation is
often attributed to the rising cost of clinical
development (2).
Longer clinical trials? Clinical trials of
drugs developed for chronic diseases take
longer than trials of occasional treatments
such as antibiotics, even if companies usually prefer to choose surrogate endpoints (cholesterol level, etc.) rather than major outcome
measures.
But the world market for cephalosporins
(the largest-selling antiinfective class), represented only one-third of the market for ulcer
treatments in 2002 (21.9 billion dollars) (3).
Slower marketing authorisation? Pharmaceutical companies often complain of slow
marketing authorisation procedures (and
other official decisions such as pricing), which
they say increase the capital opportunity costs
of research investment.
Yet, in the Tufts Center study, the drug registration period fell by 12.1 months between
1991 and 2002 (2). The International Conference on Harmonization (ICH), created at
the industry’s initiative, rapidly led to harmonisation of regulatory requirements in the
United States, Europe and Japan, leading to
cost savings and shorter development periods for drug companies. Supplementary protection certificates, which add up to 5 years
of monopoly protection after drug patent
expiry, prolong sales at premium prices (4).
More difficult marketing authorisation?
Companies often complain that regulatory
demands are increasingly difficult to satisfy.
Nevertheless, the 1980s and 1990s saw the
advent of accelerated (fast-track) marketing
authorisation procedures, mainly due to pressure from patients with AIDS. And many
marketing authorisations are now granted on
out-of-pocket spending. This simple adjustment lops about threequarters off the cost estimate published by the Tufts Center (15).
Clinical trials costs overestimated. The Tufts team set a
115 million dollar figure on mean
phase III development costs (4).
the basis of inadequate evidence, especially
in cancer and orphan diseases (see our New
Products column). And when marketing
authorisation is granted on the condition that
further trials are conducted, the data are actually provided in only a minority of cases (5).
The number of patients required to obtain
convincing results in comparative clinical trials is also used as an argument by drug companies. Yet the required group sizes increase
as the difference between treatments diminishes. In other words, comparative clinical trials are especially expensive for me-toos (a)
and other drugs that differ little from the reference drug, or that are poorly effective. In
contrast, me-toos probably cost less in terms
of preclinical research, as they are largely based
on the original drug.
Waning research? In the 1980s and 1990s,
some predicted the demise of classical pharmaceutical research based on systematic
screening of candidate compounds, claiming
that developments in biotechnology heralded a “new golden age” (6). Patients are still
waiting…
Now, the holy grail is to be found in genomics
and proteomics. It has been estimated that
genomics could cut research and development costs by 300 million dollars per new
product (7). But these hazardous predictions
are mainly intended to persuade investors to
leave one stock market bubble for another.
High-spending research. Pharmaceutical
companies, intoxicated by their multibilliondollar blockbusters of the 1990s, currently
have extremely high overheads, due to their
growing use of subcontractors, swelling intellectual property rights (royalties, etc.); and disproportionate salaries.
The death of a model?From the public health
standpoint, the sums spent on research and
development of yet another me-too product
are quite simply obscene. But companies will
continue to invest such sums if it enables them
Yet this implies either an astronomic cost per patient, or inclusion of massive numbers of
patients.
Data from the US National
Cancer Institute suggest that the
mean per-patient cost of a phase
III trial was 7 000 dollars in 19951996 (16). If we apply this unit
to cash in on a market worth ten of billions
dollars annually (anti-ulcer drugs, cholesterol-lowering agents, antidepressants, nonsteroidal antiinflammatory drugs, etc.) (4).
In contrast, these sums are ridiculously
small when it comes to finding a new shortcourse antituberculous drug (potentially saving 2 million lives each year worldwide). Nevertheless the last drug specifically developed
for tuberculosis was marketed in 1964. As a
result a public-private partnership has ben
set up to make up for the lack of industry
investment in tuberculosis treatment (8).
The key question is who really benefits
from the enormous investment in research
and development of drugs with no therapeutic advantages or no real utility? It is time for
governments to stimulate research and development oriented towards real population needs
and not simply shareholder profits. This is
the only way for health professionals and
patients to make sense of the issue of research
and development costs.
©Prescrire International
a- The term “me-too” is used for drugs with a chemical structure very similar to that of a princeps drug in a new class.
Companies develop me-toos in order to take a share of a profitable market (for example, lansoprazole, then pantoprazole, then rabeprazole, were developed in the wake of omeprazole) (ref 9).
1- Prescrire Rédaction “Le coût de la recherche pharmaceutique” Rev Prescrire 1992; 12 (123): 545-546.
2- DiMasi J et al. “The price of innovation: new estimates of drug development costs” Journal Health Economics 2003; 22: 151-185.
3-Sellers LJ “Fourth annual 50” Pharm ExecMay 2003:
42-52.
4- Prescrire Rédaction “Le renforcement tous azimuts
des brevets dans le domaine pharmaceutique” Rev Prescrire 1999; 19 (197): 544-546.
5- “FDA releases data on phase IV commitments” Scrip
2003; (2853/54): 19.
6- Sénard JM “Biotechnologies - chronique d’un succès annoncé” Rev Prescrire 1996; 16 (168): 898.
7- “A revolution in R&D -The impact of genomics”
The Boston consulting group 2001: 4 pages.
8- The global alliance for TB drug development. Website http://www.tballiance.org consulted on 4 October 2003.
9- Prescrire Rédaction “Me too” Rev Prescrire 2001; 21
(218): 404.
cost to the 115 million dollar estimate from the Tufts Center, more
than 16 000 patients would be
included in phase III trials, which
is plainly ludicrous. The Tufts
team itself calculated that an average of only 5303 patients participate in all phases of pre-marketing clinical development (4).
Even this is twice the number
(2667 patients) calculated by the
Food and Drug Administration
for 185 new chemical entities
authorised in the United States
between 1995 and 1999 (17).
Finally, regarding the special
case of orphan drugs, official US
data show that the mean P R E S C R I R E I N T E R N A T I O N A L F E B R U A R Y 2004/ V O L U M E 1 3 N ° 6 9
• 35
cost of research and development was only 8 million dollars
(including failures, and before
tax) for the 36 orphan drugs it
approved in 1998 and 1999 (18).
It was 11 million dollars overall
for the period 1998-2000, and
34 million dollars for new chemical entities (17 million after
tax) (17).
What about profits?
The cost of research and development, whatever the true figure, must be viewed in the light
of other spending, and revenues.
Billions of dollars in advertising. The pharmaceutical
industry spends about the same
amount on advertising as it does
on research and development.
French companies for example
say they devote 12.1% of their
budget to research and development and 11% to advertising
and information (19).
In 2002, five companies (Pfizer, GlaxoSmithKline, Merck,
AstraZeneca and Johnson &
Johnson) each spent more than
1.3 billion dollars on publicity
(maximum nearly three billion) (20). In 2002, six drugs
(Inexium°, Vioxx°, Tahor°,
Ogast°, Zocor° and Celebrex°)
each had a publicity budget of
more than 500 million dollars
(756 million for Inexium°) (20).
Direct-to-consumer advertising of prescription drugs cost
more than three billion dollars
in 2001, including 459 million
for GlaxoSmithKline alone (20).
Multibillion blockbusters.
In 2001, 36 drugs each had
worldwide sales of more than a
billion dollars (21). The first ten
had sales figures above 2.7 billion dollars each in 2002, Tahor°
reaching almost 8 billion (20).
The world’s biggest drug company, Pfizer, markets 10 drugs
with sales of more than a billion
dollars each (in 2002) (20).
Record profits and salaries.
For more than three decades,
pharmaceuticals have been the
most profitable industrial sector
(22). According to a study by Fortune magazine, drug company
36
profits represented 18.5% of
turnover in 2001 (compared to
12.5% for research and development). This is eight times more
than the median profitability of
the world’s 500 top companies
(all sectors), including the banking sector (13.5%), the second
most profitable (22). In 2002,
pharmaceutical companies again
outclassed all other sectors in
terms of profitability, despite their
highly publicised “difficulties” (23).
A study by economists from
Montreal university, focusing on
nine of the world’s largest drug
companies, showed that they
spent 113 billion dollars on
research and development for
the period 1991-2000, while their
shareholders reaped some 146 billion dollars in dividends (g)(24).
Drug company employees are
better-paid than workers in other
industrial sectors. A study of the
French pharmaceutical industry confirmed these findings (25).
A study of nine drug companies showed that their CEOs each
earned 21 million dollars in 2001,
excluding the value of unexercised stock options (h)(i)(26).
These high salaries obviously
impact on the cost of research
and development. For example,
doctors who recruit for clinical
trials receive several thousand
euros per patient. Thus, a British
general practice can earn 20 000
euros a year for only three hours’
work a week (27).
So, if drug research and development costs are so high, it is also
because drug companies have
very high overheads. And everyone wants a share of the pie: personnel, shareholders, researchers,
subcontractors, and health professionals (see page 35).
The real and legitimate
costs of research
and development
The Tufts Center estimate of
802 million dollars per new drug
is probably far too high.
Ideally, government authorities should themselves estimate
the cost of research and development, and especially the cost
of clinical trials. Instead, they
feebly comply with industry lob-
• P R E S C R I R E I N T E R N A T I O N A L F E B R U A R Y 2004/ V O L U M E 1 3 N ° 6 9
bies, offering international extension of patent rights, longer
patent protection, greater protection of clinical trial data and
increasingly high drug prices.
When the authorities yield to
pharmaceutical companies’
demands, especially for high drug
prices, without having access to
reliable data on research and
development costs, they ignore
the basic rules of sound economic management and risk
accelerating the demise of publicly funded welfare systems.
©Prescrire International
g- Abbott, BMS, Lilly, GlaxoWellcome,
Merck, Pfizer, Schering-Plough, SmithKline Beecham and Warner Lambert (ref
24).
h- Abbott, Allergan, BMS, Lilly, Merck,
Pfizer, Pharmacia, Schering-Plough and
Wyeth (ref 26).
i- In 2002, the (British) GlaxoSmithKline
CEO demanded an annual salary of 30
million dollars, in order to “catch up with
his American counterparts”. Shareholders
rejected this rise because it was not commensurate with the company’s stock performance. The CEO countered that he might
“feel less motivated” if he was refused the
rise (ref 29).
Selected references from Prescrire’s
literature search.
1-Prescrire Rédaction “Prix des médicaments remboursables: quelle logique?
Première partie - Années 1990 : une
libéralisation internationale croissante
défavorable au contrôle des prix des
médicaments” Rev Prescrire 2001; 21
(222): 782-786.
2-Prescrire Rédaction “Prix des médicaments remboursables : quelle logique?
Deuxième partie - Prix proportionnels
à l’innovation : principe raisonnable,
maigres résultats” Rev Prescrire 2001;
21 (223): 859-863.
3-Prescrire Rédaction “Prix des médicaments remboursables : quelle logique?
Troisième partie - Remises de fin d’année : un système discret mais peu efficace de maîtrise des dépenses” Rev
Prescrire 2002; 22 (234): 855-859.
4- DiMasi J et al. “The price of innovation: new estimates of drug development costs” Journal Health Economics 2003; 22: 151-185.
5- Tufts center for the study of drug
development. Website http://csdd.tufts.
edu consulted on 11 August 2003.
6- DiMasi J et al. “Cost of innovation
in the pharmaceutical industry” Journal Health Economics1991; 10: 107-142.
7- “Pharmaceutical R&D: costs, risks,
and rewards” Office of Technology Assessment 1993: 71 pages.
8- Public citizen “ Rx R&D myths: the
case against the drug industry’s R&D
“ scare card “”. Website http://www.citizen.org consulted on 11 August 2003.
9- Barral E “26 ans de résultats de la
recherche pharmaceutique dans le
monde (1975-2000) - Systèmes de prix
et financement de la recherche” Aventis 2001: 79 pages.
10- “Changing patterns of pharmaceutical innovation” National Institute
for Health care management 2002.
Website http://www.nihcm.org consulted on 4 October 2003.
11- “NIH contributions to pharmaceutical development - Case study
analysis of the top-selling drugs” NIH
2001: 38 pages.
12- Kaitin KI and Healy EM “The new
drug approvals of 1996, 1997 and 1998:
drug development trends in the user
fee era” Drug Information Journal 2000;
34 (1): 1-14.
13- Guenther G “Federal taxation of
the drug industry: 1990 to 1999” Congressional research service, Washington 2002: 21 pages.
14- “Large UK firms get 25 % tax credit” Scrip 2002; (2740): 3.
15- Public citizen “Tufts drug sample
is skewed; true figure of costs likely is
75 percent lower”. Website
http://www.citizen.org consulted on
11 August 2003.
16- Love J “Call for more reliable costs
data on clinical trials” Market Letter
1997: 24-25. Website http://www.
cptech.org consulted on 11 August
2003.
17-Love J “Evidence regarding research
and development investments in innovative and non-innovative medicines”
2003. Website http://www.cptech.org
consulted on 2 October 2003.
18- Love J and Palmedo M “Costs of
human use clinical trials - Surprising
evidence from the US Orphan drug act”
2001. Website http://www.cptech.org
consulted on 11 August 2003.
19- “Chiffres-clés 2002” Leem 2003:
18 pages. Website http://www.leem.org
consulted on 4 October 2003.
20-Sellers LJ “Fourth annual 50” Pharm
Exec May 2003: 42-52.
21- “2002 - the year in review” Scrip
2003; (2823): 2-3.
22- Public citizen “Pharmaceuticals
rank as most profitable industry, again”.
Website http://www.citizen.org consulted on 11 August 2003.
23-Trombella W “Industry audit &“fab
four” companies of the year”. Website
http://www.pharmexec.com consulted on 1 October 2003.
24- Lauzon LP and Hasbani M “Montants versés aux actionnaires”. In:
“Analyse socio-économique de l’industrie pharmaceutique brevetée pour la
période 1991-2000” Website http://
www.unites.uqam.ca/cese consulted
on 11 August 2003.
25- Loué JF et al. “La santé de l’industrie pharmaceutique française” Ministère de l’économie, des finances et
de l’industrie, 2002: 4 pages.
26- “Profiting from pain - Where prescription drug dollars go” Families USA
2002: 37 pages.
27- Rao JN and Sant Cassia LJ “Ethics
of undisclosed payments to doctors
recruiting patients in clinical trials” BMJ
2002; 325: 36-37.
28- “GSK does most licensing but Pfizer gets most revenues” Scrip 2003;
(2855): 14.
29- “Investors force GSK to drop pay
plan” Scrip 2002; (2803): 7.

Documentos relacionados